AI-Assisted Optical Modeling of Minimum-Visibility Routes in Urban Video Surveillance Systems

This paper presents an AI-assisted mathematical framework for identifying minimum-visibility routes in urban environments monitored by video surveillance systems. The approach is based on an analogy between intruder detection probability and light propagation in optically inhomogeneous media. Each point of the urban domain is assigned a scalar visibility coefficient derived from camera geometry and performance characteristics. The optimal path is defined as the trajectory minimizing the cumulative visibility functional. The resulting problem is reduced to the eikonal equation and solved numerically using Fast Marching and graph-based algorithms. Artificial Intelligence is employed to automatically construct the visibility field from GIS data, 3D building models, and camera metadata.